I'm beginning to wonder about one of the the DAC muxes, U2530, or the control inputs to it. The reason is that you said the trace moves vertically when the fade problem is happening. U2530, among other things, is responsible for the trace vertical position.
So, you first could take a look at U2630 pin 3 (connected to U2530 pin 2). This is the input to the opamp from the holding cap which eventually is output on U2630 pin 1. If U2630 is ok, you should see this input tracking the problem too.
Does the vertical movement happen at the same time as the fade on all the channels? Or just a particular channel or subset of channels? Can you confirm that the voltage change is happening on U2530 pins 12, 13, 14, and/or 15? These are the vertical position outputs for Ch4, Ch1, Ch2, Ch3, respectively.
U2530 has control input signals that are in common to the other DAC mux, U2521. Take a look at U2521 pin 2. See if this is moving with the fade problem. Probably not, but interesting if it is.
You could also take a look at U2530 pin 6. This is the mux inhibit signal. It probably has some periodicity to it, but for now just check to make sure the logic levels are sane (<0.7 V or >2.1V). You should also check the select inputs U2530 pins 9, 10, 11 to see if they're sane too.
Also check TP2420 (+1.36V) and TP2421 (-1.25V). These control the DAC reference voltage. From your other observations these should be stable, but let's make sure.